Patient Classification and Information Sharing for Stroke Patient Care: Insights From a U.S. Stroke Hospital
本研究利用美国一家教学医院的数据，将中风患者按病史分为四类，并建立数学模型平衡信息共享的医疗效益与成本，提出分阶段实施健康信息交换的策略，以降低死亡率并优化资源配置。
This study examines the value of patient classification and Health Information Exchange (HIE) in stroke care. Stroke care demands prompt decision-making, and this study aims to highlight the benefits of patient classification and HIE for tailored and prompt patient care. Using a dataset from a US tertiary teaching hospital, stroke patients were classified into four groups based on 22 medical history elements. A mathematical model was then developed to determine the optimal level of patient information sharing by balancing medical benefits and associated costs, thereby identifying the most appropriate level of HIE implementation. While timely access to patient medical history via HIE generally improves the effectiveness of stroke care by supporting rapid and informed decisions, a strategic approach is needed for HIE implementation. The results of this study suggest a phased HIE implementation strategy, emphasizing the trade-offs between improved patient care outcomes and HIE implementation costs. The findings demonstrate HIE's potential to reduce patient mortality and streamline care processes through effective patient classification. We also offer insights into operations management challenges in stroke patient care and emphasize the need to develop specific operating procedures and resource management policies tailored to different patient groups identified by cluster analysis. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Managerial Relevance Statement</i>-This study offers actionable insights for healthcare administrators and operations managers seeking to improve stroke care through data-driven decision making. By integrating empirical cluster analysis with analytical modeling, we demonstrate how information-sharing across a stroke network can enhance operational readiness and treatment outcomes by enabling hospitals to anticipate resource needs (e.g., ICU beds, imaging equipment, staffing) for different patient groups. Our cluster analysis highlights significant heterogeneity between stroke patients, which requires tailored resource allocation. Further, the analytical model guides decision-makers in determining the optimal level of Health Information Exchange (HIE) implementation by quantifying trade-offs between patient mortality reduction and system-level information-sharing costs. This phased approach to HIE investment provides a strategic framework for healthcare systems constrained by limited budgets. Managers can use the model to justify targeted IT investments that improve both clinical outcomes and operational efficiency. Importantly, the methodology bridges medical and engineering management perspectives, offering a replicable framework for other time-critical care processes beyond stroke treatment. Thus, our research provides healthcare leaders with tools to support data-informed patient care and its delivery. This paper also contributes to the following SDGs: SDG 3 and SDG 9.